{
  "schema": "experiment-protocol-freeze/v0",
  "experiment_id": "EXP-005",
  "title": "Completeness analyzer — Study 1 contract specification audit",
  "paper_reference": "Paper 1 · Study 7",
  "type": "heuristic",
  "frozen_at": "2026-04-15T00:00:00Z",
  "author": "Adrian Sanchez de la Sierra",
  "license": "CC BY 4.0",

  "scope": {
    "kind": "heuristic / specification audit, not a hypothesis test",
    "rationale": "This experiment does not draw new runs from a perturbation distribution. It runs a static analyzer over the traces and contracts produced by EXP-001. The output is a specification gap list, not a hypothesis test.",
    "input_corpus": "EXP-001 trace corpus (1,080 runs across refund, claim, memory)"
  },

  "method": {
    "algorithm": "For each contract C and each Study 1 trace T: enumerate consequential tool calls in T; intersect with the consequential-tools set declared in C; emit the diff. Frequency-rank the unspecified consequential tools across the full trace corpus.",
    "consequential_definition": "A tool whose call commits a state-changing effect, claim, memory write, handoff, escalation, or external publication."
  },

  "primary_question": "Are there load-bearing consequential actions present in the empirical traces but missing from the Study 1 contracts' required_tools / required_successful_order sections?",

  "secondary_question": "If yes, enumerate them so the Study 2 recall gap can be characterised as specification work (additive) rather than mechanism failure.",

  "outputs": {
    "primary": "lab/exp/exp-005-results.json",
    "input_corpus": "EXP-001 trace corpus (private to OPBR-Bench v0)",
    "code": "https://github.com/raising-agents/agent-behavior-workbench",
    "replication_cmd": "abw run --experiment EXP-005"
  },

  "claim_boundary": "This is a heuristic, not a hypothesis test. It does not establish that all future specification gaps will be enumerable in the same way. It does not establish that every consequential tool can be statically classified — consequential-status review still requires human judgment per tool. It does not test the analyzer on contracts other than Study 1's."
}